Redact Sensitive Information from Your Prompts using LLMs скачать в хорошем качестве

Redact Sensitive Information from Your Prompts using LLMs

In this video, you'll see if an LLM can act as a pseudonymization engine, allowing you to redact sensitive information from prompts before they get sent to another AI interface. You'll apply learnings from the AI Purple Teaming mini course, like in-context examples and using more than one LLM at a time, in order to evaluate whether you can appropriately find and change prompts for your use case. Notebook is here: https://github.com/kjam/secure-and-pr... This is the last video (for now) in the AI Purple Teaming mini-course, but I'm excited to hear what else you might want to learn! Check out the full mini-course here:    • Learn Hands-On AI Security & Red Teaming   If the discriminator idea is new and you want to learn more, check out the generative adversarial network (GAN) description on wikipedia: https://en.wikipedia.org/wiki/Generat... . If you've enjoyed the videos thus far, I'd appreciate a subscribe, share or like and feel free to let me know what you've learned or would like to learn in the comments.